KonCite · Investigative Public Intelligence

Before the Gate Closes: Black Ownership in the Age of AI

Black labor helped move America through the horse and automobile economies while control of the most valuable routes consolidated elsewhere. Artificial intelligence has opened another road. The question is whether Black communities will merely use it—or own the knowledge, applications, standards, contracts, and distribution systems through which the new economy will move.

Kevin Ahmaad Jenkins, PhD Technology, Labor & Ownership 22–26 minute read

In the spring of 1905, Chicago’s freight economy began to seize. Teamsters stopped moving goods from railway depots and warehouses, employers recruited replacement drivers, police entered the streets, and businesses discovered how quickly ownership became meaningless when no one could move what they owned. Among the organized drivers were roughly 2,000 Black Teamsters—men whose presence complicated every easy story about race, labor, solidarity, and power in the early union.

The word teamster originally named a person who drove a team of horses. The title sounded ordinary because the horse and wagon had become ordinary, but the work sat at the center of urban commerce. Teamsters carried coal toward furnaces, food toward markets, beer toward taverns, furniture toward homes, merchandise toward department stores, and raw materials toward factories. They rarely owned the freight, and many did not own the animals, wagons, or commercial accounts that made each trip profitable. Yet the transaction failed without them.

A teamster needed enough strength to control a loaded wagon and enough intelligence to coordinate weight, terrain, animal endurance, weather, congestion, customer expectations, and the unwritten geography of loading docks and alley entrances. The horse supplied propulsion. The driver supplied judgment. The route supplied economic power.

By 1900, the United States counted more than half a million teamsters. Labor historian David Witwer’s analysis of census records found that Black men represented approximately 12.5% of that national workforce. Across eight Southern cities with populations above 50,000, Black workers constituted, on average, nearly three-quarters of the teamster workforce.1

The governing question is therefore not whether Black people entered technological revolutions. They did. The governing question is: when the engine changes, who inherits control of the route?

The history of theTeamsters reveals that technological revolutions do not automatically redistribute power. The horse disappeared, the truck arrived, and Black workers remained essential to transportation; yet employers, unions, financiers, carriers, and property owners continued to determine who received the most valuable routes. Artificial intelligence now creates a comparable formation period. Black people already use the technology, contribute to the culture, and understand many of the institutions it promises to transform. The unresolved question concerns whether Black communities will own the data, applications, standards, contracts, and distribution systems through which the new economy will move.

The governing finding

The central racial risk of artificial intelligence is not Black nonparticipation. It is Black participation without control of the valuable routes.

Transportation history shows how a community can supply labor, expertise, culture, and consumer demand while ownership consolidates elsewhere. Artificial intelligence creates another contest over models, data rights, institutional contracts, cultural attribution, standards, distribution, and recurring economic value.

How movement became union power

At the beginning of the twentieth century, teamsters often worked 12 to 18 hours a day, seven days a week, for approximately $2 a day. Employers could hold drivers responsible for damaged goods, lost merchandise, or unpaid customer accounts. The driver carried commercial risk without corresponding authority over the business that created it.

In 1903, two driver organizations merged to create the International Brotherhood of Teamsters. The union organized horse-team drivers, helpers, stable workers, and others who occupied the final commercial passage between producers and customers. It learned quickly that the teamster’s leverage did not reside only in the animal. The leverage came from the dependency surrounding the route. A factory could own machinery, a merchant could own inventory, and a railroad could carry freight into the city, but none of those assets could complete the exchange when local movement stopped.

Black teamsters participated in the union from its formative years. Black delegates attended the founding convention; some locals organized across racial lines; and some Black members held office. One Black Philadelphia delegate described a local of approximately 200 members, split nearly evenly between Black and White workers, that had elected him president. Historians estimate that Black membership may have reached roughly 6,000 out of slightly more than 42,000 Teamsters by 1912.

Those facts prevent a crude claim that the Teamsters formed as a uniformly White racial monopoly. The early organization sometimes pursued Black membership more aggressively than other craft unions because leaders understood that employers could use excluded Black drivers to weaken wages and strikes. Yet the absence of a national racial ban did not create equal local power. The union’s decentralized structure allowed local officers and employers to shape membership, seniority, assignments, and access to protected work. Some locals integrated. Others segregated. Some Black members attained office. Others found the premium gate farther down the road.2

Table 1

Black Participation Did Not Guarantee Control of the Route

The historical record documents meaningful Black participation in transportation labor and union activity. It does not, by itself, establish equal access to ownership, premium assignments, wages, institutional authority, or the assets surrounding the route.

Evidence showing how Black participation in transportation did not automatically produce control over routes, contracts, ownership, or institutional authority.
Historical evidence	What it establishes	What it does not establish
Black men represented approximately 12.5% of US teamsters in 1900.	Black workers held a substantial position inside the horse-era distribution economy and helped move goods through cities, markets, freight yards, businesses, and households.	The figure does not establish equal wages, horse or wagon ownership, freight-contract control, route quality, union power, or commercial-property access.
Some early Teamsters locals included Black members, elected Black officers, and organized across racial lines.	The early union did not maintain one universal national rule of total racial exclusion, and Black workers participated in its institutional formation.	Integrated examples do not prove consistent local equality, equal leadership authority, or uniform access across cities, regions, employers, and freight categories.
Federal investigators documented union-employer barriers that restricted Black drivers from over-the-road trucking in Detroit.	Closed-shop arrangements and local gatekeeping could control access to premium routes, protected jobs, stronger wages, and more desirable freight.	The Detroit case does not prove that every Teamsters local or every trucking market used identical discriminatory practices.
Black workers remain highly represented in transportation and several driving occupations.	Black labor continues to operate, sustain, and move the modern transportation system at rates above the overall workforce benchmark in several occupations.	Occupational representation does not measure fleet ownership, contract control, route profitability, equity, capital access, logistics-platform ownership, or authority over how work gets allocated.

America changed the engine

The automobile did not replace the horse through a single invention or a clean national choice. America rebuilt the economic environment around mechanical mobility over several decades. In 1910, fewer than 500,000 motor vehicles operated in the United States. By 1920, registrations approached 10 million. By 1930, they exceeded 26 million.3

The transformation reached far beyond automobile factories. Feed became fuel. Stables became garages, depots, and freight terminals. Farriery and wagon repair gave way to tire service, mechanical repair, and body work. Insurance expanded from animal mortality and carriage loss into collision, liability, theft, and commercial fleet risk. Watering and feeding networks gave way to service stations and petroleum distribution. Local horse delivery stretched into regional and interstate freight.

The Teamster survived because his institutional power never depended completely on the horse. The animal supplied force, but the driver controlled circulation. When the horse vanished from the commercial center, the union released the reins and took the steering wheel. The wagon became the truck; the stable became the terminal; the feed contract became the fuel contract; the freight remained.

Technological transition often hides this continuity. Society celebrates the visible machine while overlooking the institutions that carry authority from one era into the next. America modernized the vehicle without automatically modernizing the gate.

Figure 1

The Vehicle Changed, but the Route Remained

Across three technological eras, the visible machine changed while the central economic struggle remained remarkably stable: who controlled movement, access, contracts, and the assets that generated the greatest return.

Horse economy

Living horsepower

Vehicle: Horse-drawn wagon
Freight: Physical goods and passengers
Operator: Teamster
Gatekeepers: Merchant, employer, union, and property owner

Premium asset

Route, freight contract, stable access, and commercial account

Automobile economy

Mechanical horsepower

Vehicle: Motor truck
Freight: Regional and national goods movement
Operator: Truck driver
Gatekeepers: Carrier, union, financier, regulator, and fleet owner

Premium asset

Premium route, fleet ownership, contract, and seniority

AI economy

Intelligence power

Vehicle: Model, application, and platform
Freight: Information, judgment, culture, and workflow
Operator: User, implementer, developer, and institutional buyer
Gatekeepers: Platform owner, investor, procurement office, and data owner

Premium asset

Data rights, model, institutional contract, standard, and distribution

The analogy concerns control over movement, not technological equivalence. The horse, truck, and AI platform perform different work, but each era creates valuable routes through which goods, services, judgment, and economic power travel.

Different freight, different future

Black workers followed transportation from horse-drawn wagons into trucks, taxis, buses, delivery fleets, warehouses, garages, and automobile plants. The new technology expanded opportunity, but employers and unions still assigned value through racial categories.

A federal Fair Employment Practice Committee investigation documented the mechanism in Detroit during the 1940s. Teamsters Local 299 held closed-shop agreements with major trucking companies and denied Black drivers access to membership and over-the-road jobs. At the same time, Black drivers regularly hauled ashes, garbage, coal, and furniture—work the federal report described as heavier, dirtier, and less remunerative than protected long-distance freight.4

The discrimination did not claim that Black men lacked the capacity to drive. It assigned them a different freight.

That arrangement created two transportation economies inside the same industry. Black drivers carried undesirable loads through local routes with weaker protection and limited advancement, while established networks reserved premium cargo, seniority, union security, and better compensation for White workers. Both groups appeared in transportation. Only one received the full economic benefit of belonging.

This is the more precise meaning of racialized route control. The institution did not need to exclude every Black driver from every road. It needed to govern the gateway to the routes that paid best.

Black freight driver standing confidently beside an early commercial motor truck loaded with cargo outside a transfer company. — **Black labor in the early motor-freight economy.** Black workers entered motorized transportation as drivers, mechanics, freight handlers, chauffeurs, delivery workers, and factory laborers. Their movement into the new technology expanded participation, but it did not automatically transfer ownership of fleets, premium routes, freight contracts, finance, or institutional authority.

Image note

Editorial historical reconstruction created for *Who Owns the Route?*

This image is an editorial reconstruction rather than a verified archival photograph. It should not be assigned a historical date, named subject, company history, or documentary archive citation.

Black mobility built its own intelligence layer

The automobile era did more than reorganize Black labor. It changed Black mobility. A privately owned car could reduce dependence on segregated streetcars, railroads, buses, and local schedules. It gave families greater control over departure, destination, companions, and the private space of the journey. Yet the road did not erase segregation. It relocated racial danger into hotels, restaurants, service stations, repair shops, police encounters, and towns where a Black motorist could not assume that money guaranteed service.

Victor Hugo Green answered that problem by publishing The Negro Motorist Green Book beginning in 1936. The guide compiled lodging, restaurants, service stations, tourist homes, and other establishments where Black travelers could seek service with greater confidence. It did not manufacture the automobile or pave the highway. It built an intelligence layer on top of the dominant mobility platform.5

That distinction changes the AI analogy. Black communities have already built systems that convert lived knowledge into navigation, trust, safety, and commerce. The Green Book organized distributed information, verified places of refuge, redirected Black consumer spending, and made a hostile infrastructure more usable without pretending that the infrastructure had become fair.

AI creates a similar opening. Black institutions do not need to own every foundation model to build the cultural, administrative, and institutional intelligence that makes general technology useful in Black life.

Cover or directory page from The Negro Motorist Green Book, a travel guide created for Black motorists during segregation. — **The Negro Motorist Green Book** converted collective Black travel knowledge into practical infrastructure by identifying businesses where Black motorists could seek service with greater confidence. The guide did more than list destinations. It helped Black travelers navigate risk, locate welcome, and connect mobility to Black-owned commerce.

Credit and link

Credit: Smithsonian or holding archive. View the archival record

Replace **PASTE-ARCHIVE-RECORD-URL-HERE** with the direct Smithsonian or holding-archive record URL for the specific cover or interior page you are citing.

Artificial intelligence carries interpretation

Artificial intelligence does not move coal, beer, or furniture. It moves interpretation. Its freight includes language, research, medical information, legal records, employment decisions, educational content, institutional memory, images, voices, customer behavior, administrative judgment, and cultural meaning.

Its roads consist of models, datasets, cloud systems, application interfaces, software platforms, procurement contracts, security protocols, and professional standards. Its drivers include the people who prompt systems, review outputs, correct errors, implement tools, train colleagues, label data, and redesign workflows. Its gatekeepers include the companies and institutions that own models, computing infrastructure, proprietary data, customer relationships, capital, standards, licenses, and enterprise agreements.

Black people have already entered this economy. A 2025 Jobs for the Future survey reported that 83% of Black respondents were familiar with AI, 53% used it daily or weekly, 67% said it already affected their work, more than 80% expected career effects within three to five years, and 71% believed they needed additional skills. The sample does not function as a federal labor census, but it undermines the assumption that Black communities stand outside the technology waiting for an invitation.6

Black users recognize practical leverage. AI can help a patient organize years of symptoms before a medical appointment, a parent interpret special-education records, a business owner decode a procurement notice, a scholar search an archive, a church preserve institutional memory, and a family organize legal or financial documents that no one else has time to explain. The automobile expanded the radius of physical mobility. AI can expand the radius of institutional mobility.

Table 2

Black Presence in Transportation and Technology Does Not Automatically Produce Control

Black workers remain highly represented in several transportation occupations while representation across key technology occupations is uneven. Neither pattern, standing alone, tells us who owns the fleet, controls the platform, holds the contract, governs the data, or captures the recurring value.

Transportation and logistics Technology and computing

Black workforce representation in selected transportation, logistics, and technology occupations in 2025, compared with the Black share of total employment.
Occupation or workforce category	Black share	Workforce benchmark	Analytical meaning
Total US employment	12.7%	12.7%	National workforce benchmark used to interpret concentration above or below the overall Black share of employment.
Transportation and material moving occupations	20.0%	12.7%	Black workers hold a substantially above-benchmark presence across the broad occupational system that moves people and goods.
Driver/sales workers and truck drivers	20.2%	12.7%	Strong operational presence behind the wheel does not reveal who owns the vehicle, route, carrier, brokerage relationship, or freight contract.
Transit and intercity bus drivers	32.2%	12.7%	Very high representation in system operation does not establish authority over transit policy, capital investment, contracts, scheduling systems, or ownership.
Transportation, storage, and distribution managers	13.4%	12.7%	Representation is close to the national workforce benchmark, but the category still does not measure equity ownership, procurement authority, capital access, or corporate control.
Computer and mathematical occupations	9.5%	12.7%	Black representation falls below the workforce benchmark across the broad occupational category that includes many technical roles shaping digital systems.
Software developers	5.4%	12.7%	Significant underrepresentation in a role central to building applications and platforms, though employment share still does not measure company ownership or product authority.
Computer programmers	6.6%	12.7%	Representation remains well below the national benchmark in an occupation responsible for translating specifications into functioning code.
Computer support specialists	13.1%	12.7%	Representation is near the workforce benchmark in a role that supports system operation, but operating and maintaining a system is not the same as owning or governing it.
Database administrators and architects	6.5%	12.7%	Underrepresentation appears in a role closely connected to data architecture, storage, access, and the organization of institutional knowledge.
Network and computer systems administrators	17.3%	12.7%	Above-benchmark operational representation demonstrates that technical participation varies by role, but employment still does not establish ownership of infrastructure or standards.

The gate forms before it looks closed

The emerging gate does not need a sign that prohibits Black entry. It can operate through accumulated requirements that appear neutral when examined one at a time: elite credentials, investor introductions, expensive computing, unpaid time to build a portfolio, enterprise references, cybersecurity certifications, insurance thresholds, long procurement cycles, and access to proprietary data.

These requirements do not prove intentional racial discrimination in every case. They do, however, distribute opportunity through markets that begin with unequal capital, networks, institutional sponsorship, and property. A founder who understands the customer but cannot survive an eighteen-month sales cycle may lose to a less knowledgeable competitor with deeper financing. A Black institution may supply valuable workflows and community trust while a vendor retains the data, software, license, and recurring revenue.

The old closed shop required union membership. The emerging AI closed shop may require compute, capital, data, contracts, certifications, and someone inside the network willing to open the door.

The ownership problem therefore cannot collapse into a demand for more Black coders. Technical talent matters, but the route also passes through procurement officers, clinicians, lawyers, archivists, linguists, insurers, investors, educators, cultural institutions, and distribution networks. The economy will reward those who organize these assets into systems that customers cannot easily replace.

Figure 2

The Participation-to-Power Route

A community can gain access to a technology, use it extensively, and become highly visible inside its workforce without controlling the assets, rules, contracts, institutions, or economic returns that make the system powerful.

The route begins with presence. It reaches power only when participation develops into authority, governance, ownership, wealth creation, and the ability to set standards for others.

Stage 01

Access

The technology, institution, market, or opportunity becomes available to enter.
Entry
Stage 02

Use

People adopt the system, depend on it, generate activity, and create demand.
Adoption
Stage 03

Participation

People work inside the system, contribute labor, knowledge, culture, data, and expertise.
Contribution
Stage 04

Representation

Presence becomes visible in employment, leadership pipelines, public narratives, and institutional reporting.
Visibility
Stage 05

Authority

People gain the recognized ability to make consequential decisions rather than merely advise those who do.
Decision rights

Governance
Establishing rules, protections, permissions, accountability, and institutional boundaries.
Control
Determining how assets, routes, data, contracts, and opportunities are allocated.
Ownership
Holding the enterprise, platform, intellectual property, customer relationship, or underlying asset.
Wealth Creation
Retaining recurring value, equity, licensing income, appreciation, and intergenerational benefit.
Standard Setting
Defining what qualifies, what gets measured, who may participate, and which terms govern the market.

Representation is not the destination. The strategic question is whether Black communities can move from using and staffing artificial intelligence to governing its data, owning its applications, controlling institutional contracts, retaining the value it creates, and setting the standards by which systems enter Black institutions.

AI can enhance the Black experience

Administrative mobility

AI can reduce the informational advantage that large institutions hold over individuals. It can help people interpret letters, assemble timelines, identify deadlines, compare policies, organize medical histories, prepare questions, and recognize when professional assistance has become necessary.

Institutional mobility

Black churches, HBCUs, clinics, media companies, nonprofits, and small businesses often possess vision and trust but operate with thin administrative capacity. Secure AI systems can improve research, compliance, documentation, fundraising, procurement, scheduling, and knowledge retention without replacing human judgment.

Cultural mobility

Black history remains dispersed across newspapers, oral histories, church records, funeral programs, photographs, scholarship, music, and family archives. AI can help make those materials searchable, trace intellectual lineage, preserve provenance, and return attribution to people and institutions whose work commercial systems often detach from its origins.

Economic mobility

The largest opportunity comes when communities turn domain knowledge into owned products, governed datasets, licensing systems, certification standards, specialized applications, and recurring contracts. Lived experience becomes economically powerful only after someone packages, protects, and governs it.

AI must learn Black life in full resolution

Black life cannot enter AI primarily through police records, hospital records, disciplinary files, credit denials, and datasets organized around disparity. Those sources document real institutional encounters, but they teach a system to recognize Black people most clearly when something has gone wrong.

A serious Black cultural intelligence infrastructure would include Black newspapers, HBCU archives, church records, oral histories, scientific scholarship, business histories, local journalism, regional language, family records, intellectual traditions, artistic lineages, and the ordinary documentation of Black life. Models need context, not a racial glossary.

Research has shown that hate-speech detection systems and human annotators can associate features of African American English with toxicity or offensiveness. The failure does not originate in the language. It originates in the labels, assumptions, and defaults that teach the system which expressions count as normal.8

Cultural understanding without attribution can become a more efficient form of extraction. A model should not absorb generations of Black language, scholarship, sermons, images, music, and community knowledge and then present the resulting intelligence as though it emerged without authorship. Provenance, licensing, consent, permitted use, attribution, and compensation belong inside the technical architecture.

Ownership Framework

The Four Layers Black Institutions Can Own

Black institutions do not need to manufacture every foundation model to exercise meaningful power in artificial intelligence. They can own the knowledge, applications, assurance systems, and market channels that determine how AI enters Black life and where its value accumulates.

Each layer represents a distinct place where ownership, control, licensing, recurring revenue, institutional authority, and cultural protection can be built. Together, they form an economy rather than a collection of disconnected tools.

Knowledge Layer

Own the intelligence AI depends upon

Black institutions can govern the archives, datasets, language, professional knowledge, cultural records, and community memory that make artificial intelligence useful and trustworthy.

Permissioned Black archives and knowledge repositories
Cultural attribution and provenance systems
Institutional memory and family-history infrastructure
Community-owned datasets with explicit consent terms
Licensing standards for cultural and scholarly knowledge

Strategic value: control over source material, permission, attribution, access, licensing, and reuse.

Application Layer

Own the systems that solve institutional problems

Black institutions can build or co-own specialized AI tools that improve how people navigate healthcare, education, law, business, faith, media, research, and public systems.

Healthcare preparation and patient-navigation systems
Education, advising, and credential-mobility platforms
Legal and public-benefit navigation tools
Black business operations and procurement systems
Research, church, media, and nonprofit workflow tools

Strategic value: ownership of the workflow, customer relationship, institutional contract, subscription revenue, and product roadmap.

Assurance Layer

Own the systems that determine whether AI qualifies

Black institutions can define how systems are tested, audited, secured, certified, insured, and approved before they enter communities or high-stakes institutional environments.

Cultural competence and language-performance testing
Algorithmic bias and outcomes auditing
Cybersecurity and privacy assurance
Risk classification and institutional readiness review
Certification, insurance, and quality-control standards

Strategic value: authority to determine what is safe, credible, culturally responsive, compliant, and institutionally fit.

Market-Control Layer

Own the routes through which AI reaches institutions

Black institutions can aggregate demand, negotiate terms, direct procurement, control distribution, finance products, and establish the commercial standards governing access to their markets.

Purchasing and procurement consortiums
Institutional distribution and customer networks
Licensing, credentialing, and approved-vendor systems
Black investment vehicles and product-financing structures
Shared contracting, negotiation, and data-governance terms

Strategic value: bargaining power, market access, contract control, capital formation, distribution, and recurring economic return.

Knowledge
determines what the system knows

Applications
determine what the system does

Assurance
determines whether the system qualifies

Market control
determines who captures the value

The strategic objective is not simply more Black AI users. It is a Black institutional economy in which communities can own the knowledge, govern the applications, certify the systems, negotiate the contracts, control distribution, and retain a meaningful share of the value created.

Becoming the gatekeepers

The word gatekeeper carries the memory of exclusion, but every functioning system contains gates. Someone decides which information enters, who may use it, what qualifies as evidence, which standards apply, what requires consent, and who receives compensation.

Black institutions can govern gates without rebuilding the racial hiring hall. A legitimate gate can require consent before a company trains on an archive, cultural validation before a hospital deploys a model, attribution before a system reproduces a scholar’s work, security before a vendor accesses community data, and compensation before a company synthesizes a creator’s voice or likeness.

That posture differs fundamentally from asking a dominant platform for better representation. It establishes the conditions under which the platform may enter.

Before Black institutions adopt an AI system, ask:

Who owns the customer relationship?
Who owns the data created through use?
Can the vendor train on institutional or community information?
Who owns the workflow and intellectual property?
Can the institution leave with its data and operational knowledge?
Who verifies cultural and technical performance?
Does recurring value accumulate inside or outside the institution?
Can Black institutions jointly purchase, license, or co-own the system?

The road remains under construction

The Black teamster often knew the route better than the merchant who owned the freight. The Black truck driver could master the machine while unions, carriers, banks, and fleet owners controlled access to the most valuable miles. Artificial intelligence gives Black America a chance to interrupt that sequence before it hardens into another industrial inheritance.

The opportunity does not require us to manufacture every foundation model or own every data center. It requires us to recognize the assets already in our hands: trusted institutions, cultural knowledge, professional expertise, archives, language, workflows, distribution networks, community legitimacy, and lived knowledge of how American systems actually behave.

The gatekeepers have started building. They have not finished.

This time, Black people can own more than the vehicle that carries the future. We can own the knowledge it depends upon, the standards it must satisfy, the routes through which it enters our institutions, and the terms under which it creates value.

The Teamster survived because the route mattered more than the horse.

The road beneath artificial intelligence remains under construction. Black America should help decide where it leads—and collect the toll when others travel through what we built.

Evidence Record

Sources and Notes

Open each entry to review the complete citation, original source, interpretive limitation, and evidentiary boundaries governing its use in this investigation.

Peer-reviewed historical scholarship Race Relations in the Early Teamsters Union

Witwer D. Race relations in the early Teamsters Union. Labor History. 2002;43(4):505-532.

Read the historical labor study (PDF)

Limitation: Historical census and union evidence. Counts reflect source coverage and terminology of the period.

Institutional history International Brotherhood of Teamsters: The Early Years

International Brotherhood of Teamsters. The Early Years. Accessed July 10, 2026.

Review the Teamsters institutional history

Limitation: Union-produced history; use for institutional chronology and stated working conditions, not independent evaluation.

Federal historical record From Names to Numbers: The Origins of the US Numbered Highway System

Weingroff RF. From Names to Numbers: The Origins of the US Numbered Highway System. Federal Highway Administration.

Review the FHWA historical record

Limitation: Official federal historical overview; vehicle counts and road history should be read within the source’s coverage.

Federal investigation President’s Committee on Fair Employment Practice Final Report

President’s Committee on Fair Employment Practice. Final Report. US Government Printing Office; 1946.

Read the federal fair-employment report

Limitation: The Detroit trucking finding documents a specific institutional case and should not be generalized to every local.

Museum and archival interpretation About The Negro Motorist Green Book

Smithsonian National Museum of African American History and Culture. About The Negro Motorist Green Book.

Explore the Smithsonian Green Book project

Limitation: Institutional interpretation of the guide’s history and function.

Workforce survey and strategic report AI for Black Learners and Workers: An Equity Roadmap

Juncos A, Collins M, Swartsel A, et al. AI for Black Learners and Workers: An Equity Roadmap. Jobs for the Future. 2025.

Read the AI equity roadmap

Limitation: Survey findings describe the reported sample and do not function as a federal labor-market census.

Federal workforce data Employed People by Detailed Occupation, Race, and Ethnicity

US Bureau of Labor Statistics. Employed people by detailed occupation, sex, race, and Hispanic or Latino ethnicity, 2025 annual averages.

Review the official BLS occupation table

Limitation: Occupational shares measure workforce representation, not ownership, compensation, equity, route quality, or contract control.

Peer-reviewed AI research The Risk of Racial Bias in Hate-Speech Detection

Sap M, Card D, Gabriel S, Choi Y, Smith NA. The risk of racial bias in hate speech detection. In: Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. 2019:1668-1678. doi:10.18653/v1/P19-1163.

Read the peer-reviewed conference paper

Limitation: The findings concern hate-speech detection and annotation; they should not be generalized to every AI system.

Federal STEM workforce evidence The STEM Labor Force

National Center for Science and Engineering Statistics. The STEM Labor Force: Scientists, Engineers, and Skilled Technical Workers. National Science Foundation; 2024.

Review the NSF STEM workforce evidence

Limitation: Representation varies by occupation, degree, workforce category, and analytic definition.

Federal risk-management framework Artificial Intelligence Risk Management Framework

National Institute of Standards and Technology. Artificial Intelligence Risk Management Framework.

Review the NIST AI risk framework

Limitation: Framework guidance; not a finding that any particular system meets or fails the standard.

Before the Gate Closes: Black Ownership in the Age of AI

Before the Gate Closes: Black Ownership in the Age of AI

The central racial risk of artificial intelligence is not Black nonparticipation. It is Black participation without control of the valuable routes.

How movement became union power

America changed the engine

Living horsepower

Mechanical horsepower

Intelligence power

Different freight, different future

Black mobility built its own intelligence layer

Artificial intelligence carries interpretation

The gate forms before it looks closed

Access

Use

Participation

Representation

Authority

AI can enhance the Black experience

Administrative mobility

Institutional mobility

Cultural mobility

Economic mobility

AI must learn Black life in full resolution

Own the intelligence AI depends upon

Own the systems that solve institutional problems

Own the systems that determine whether AI qualifies

Own the routes through which AI reaches institutions

Becoming the gatekeepers

Before Black institutions adopt an AI system, ask:

The road remains under construction

Who owns the black digital twin?